High-strength steel becomes standard on Alberta gas system

Use of Grade 550 (X-80) pipeline steels in 1994 and 1997 along the gas-pipeline system operated in Alberta by NOVA Gas Transmission (NGT) led to material savings, greater gas-flow capacity, and increased fuel-gas savings.

As a result, the material is now the basis of a standard platform for the design and construction of large-diameter pipeline projects in the network.

Before its merger with TransCanada PipeLines Ltd. (TCPL) earlier this year (box), NGT owned and operated more than 22,000 km of sweet natural-gas pipelines in Alberta. In 1997, the pipeline system transported 80% of the natural gas produced in Canada, equaling 15% of all natural gas produced annually in North America.

One of the most complex and technologically advanced systems in the world, the system has been evolving over the past 40 years, resulting in a diverse network spread over a broad geographic area.

The specified minimum yield strength (SMYS) of the pipe in the NGT system has increased steadily over this time as NGT sought to respond to the changing needs of customers to provide gas transportation at the lowest possible life-cycle cost.

Material introduction

NGT estimates costs saved by its customers as a result of introducing Grade 483 steels alone at more than $200 million. Use of this technology together with other cost-effective design and construction techniques have enabled NGT to install large-diameter facilities at the lowest cost in North America (OGJ, Nov. 25, 1996, p. 39; Aug. 4, 1997, p. 37).

Key to the success of the introduction of Grade 483 steels for large-diameter pipeline construction were the availability of a Canadian supply, the establishment of mechanized welding as the preferred process for main line welding, and a detailed understanding of pipe steel and weld behavior and the effect of weld-metal strength and toughness on pipeline integrity.

The success of using Grade 483 pipe materials led NGT in the early 1990s to consider application of Grade 550 material; initial analyses of the benefits of using higher strength pipe are shown in Fig. 1 [28,949 bytes]. A large number of combined technologies were required to implement Grade 550 successfully; the following sections describe how that success was achieved.

Canadian supply

The composition was based on Ipsco's Grade 483 chemistry, which is low-carbon, manganese-niobium, with molybdenum and titanium additions. To increase the strength while maintaining weldability and toughness, the manganese was raised only slightly and the required properties achieved through combination of a further addition of micro-alloys along with a two-stage, thermo-mechanical controlled rolling process.

Extensive understanding of the complex precipitation and controlled rolling parameters led to successful development of the Grade 550 material. A parallel program that examined mechanical properties of the pipe to determine the effects on both pipeline construction and in-service characteristics accompanied this development.

Additionally, it was important to ensure that pipeline contractors would not view application of Grade 550 as difficult. NGT worked with contractors to demonstrate field construction behavior, including bending and welding, leading ultimately to a successful project in 1994.

Welland Pipe Ltd., part of Stelco Inc., Hamilton, Ont., developed Grade 550 in another co-operative effort between industry, Canadian research laboratories, and universities. Welland Pipe used a manganese-niobium-titanium alloy system with additions of nickel and molybdenum.

Controlled rolling followed by accelerated cooling was used to produce a fine bristle-like ferrite microstructure and the required strength, toughness, and weldability. Development was followed by a series of successful NGT projects in 1997.

Historically for large-diameter pipeline construction, NGT has used two approaches to main line welding, mechanized GMAW or manual shielded metal arc welding (SMAW).

The approach adopted would depend on economic considerations: the most cost-effective use of mechanized and manual welding depends on the type of mechanized welding system, the length of each individual construction project, and the topography of the land to be traversed.

Shorter projects, less than approximately 25 km, were typically built with manual SMAW, although this is now rapidly changing. These projects use radiographic inspection and follow standard workmanship rules for acceptance of weld defects.

For most projects longer than 25 km or for higher-strength pipe, mechanized GMAW is used in conjunction with mechanized ultrasonic testing (UT) inspection and an alternative defect-acceptance standard based on engineering critical assessment (ECA).

Manual welding

Repairs are made with a combination of E5510G, vertical down, for the root and hot pass, and E55018C2 low-hydrogen electrodes, vertical up, for fill and cap passes.

As discussed presently, some of these welds can be slightly undermatched in strength level (Fig. 2 [34,533 bytes]), but this has not been an issue because the welds have always had sufficient toughness. Minimum Charpy energy for these welds is generally greater than 30 ft-lb at -5° C.

Previously, NGT has relied on the inherent conservatism of "workmanship"-based codes to ensure integrity of Grade 483 SMAW welds. Brittle fracture is avoided through the control of allowable defect size in these tough welds, thus ensuring that the failure is controlled by plastic collapse.

Most SMAW defects are non-planar defects such as undercut, hollow bead, and porosity, and their distribution and size are accounted for in the assessment.

For main line welding of Grade 550, manual SMAW with cellulosic electrodes is not considered an option. Welding trials by NGT determined that the desired combination of strength and toughness is difficult to achieve and the higher-strength cellulosic electrodes, especially those used for the root pass, are less than ideal for high productivity applications.

In addition, the likelihood of hydrogen-assisted cracking is increased as a result of the higher restraint and increased level of strain that a root bead will experience as a result of both the higher-strength steel and higher-strength weld metal. For tie-in welds and repairs, procedures have been developed which aim to achieve high yield strength, yet retain good weldability.

SMAW E55010 electrodes are used for both the root and hot passes. This has proven to be the best approach to maintain handling characteristics; good root-bead handling is important both to avoid defects and to create a bead profile resistant to hydrogen cracking.

Fill and cap passes are completed with either the self-shielded, flux-cored arc welding (FCAW) process with an E9IT8-G wire used for tie-in welds, or the low-hydrogen, vertical-down SMAW process with E69018G electrodes used for main line applications.

These welds may have slightly more under-matching than manual main line welds on Grade 483 (Fig. 2 and Fig. 3 [34,753 bytes]).

Mechanized welding

Two types of mechanized-welding systems have been used. In both, the majority of the external weld passes are deposited with lightweight tractors ("bugs") running on a band to carry the welding head around the pipe.

The systems use small-diameter wires at relatively high current to give high metal-deposition rates, vertical-down weld progression, and are deposited in a reduced-gap, compound bevel that is accurately machined on the pipe ends immediately ahead of the welding crew.

The major difference between the systems is in the deposition of the root bead: one uses welding heads incorporated into the internal line-up clamp to produce a root bead from inside the pipe; the other system completes all passes externally, with the root bead being run on to a copper backing bar which is incorporated into the internal line-up clamp.

To date, both systems have used Thyssen K-NOVA wire for the Grade 550 construction projects, although other wires are under consideration. Yield strength distributions for the mechanized GMAW process and K-NOVA wire are included on Figs. 2 and 3.

Mechanized GMAW is a low-hydrogen welding process and has several advantages over manual processes with respect to weld-metal strength mismatch. Strength and toughness are consistent in mechanized GMAW welds. Also, the welds are much narrower, and therefore any under-matched welds benefit from the additional constraint.

The process can exhibit a tendency towards lack-of-fusion defects; these, however, are consistently and reliably detected with mechanized ultrasonic testing systems. A key result, due to the repeatability of mechanized GMAW, is that it is easily matched to a mechanized UT system.

UT inspection offers an advantage because both defect length and depth are obtained, thus allowing alternative weld-defect acceptance criterion to be developed. Repairs to mechanized gas-metal-arc welds are performed with manual SMAW and any defects in a repair welds are accepted to the workmanship standard.

Mismatch; acceptance criteria

This inattention, however, did not lead to any serious problems. This has been because generally the welding consumables overmatched the pipe yield strengths and primarily because the pipe has been of lower yield strength. As a result, this led to the belief that overmatching was good and undermatching bad.

This philosophy, however, can lead to some undesirable results when applied to higher-strength pipe materials, especially Grade 550. As shown in Figs. 2 and 3, to achieve consistent overmatching in higher-strength materials would lead to the application of extremely high-strength weld metals, with their accompanying problems with respect to weld hydrogen cracking and lower toughness.⁷

NGT's approach to this issue was to understand the weld design fully and its impact on the pipeline design, thereby achieving a balance between weldability and weld performance.

Two approaches can be used for design of a pipeline, stress or strain based.

The conventional approach, stress based, has been used extensively worldwide to build safe and reliable pipelines. The stress-based approach relies essentially on limiting the applied hoop stress to lower than the yield stress.

Strain-based design of pipelines is most useful when the anticipated longitudinal strains are likely to cause plastic deformation in either the pipe or the weld metal. These conditions most often arise under conditions of ground movement, such as subsidence or upheaval by permafrost.

In either case, the applied longitudinal stress/strain is the factor that must be considered for girth-weld defect assessment.

Strain-based design requires careful consideration of the interaction between applied strain, material toughness, and tolerable defect size, factors not covered here.

Stress-based design

The yield strength for pipe materials is usually specified at a total strain under load of 0.5%. Under these nominally elastic conditions, the well-recognized methods for determining alternative acceptance criteria, based on both fracture and plastic collapse, can be used to determine acceptable defect sizes.

The actual failure mode at a girth-weld defect will depend on driving forces and the resistance of that particular defect. The driving force that should be considered is the highest longitudinal stress that will act on the defect.

For a buried pipeline operating at a hoop stress of o80% SMYS, this is often the lowering-in stress during construction. The resistance of a defect depends on the size, wall position and circumferential location, and the strength and toughness of the material surrounding the defect.

Most fracture-based assessments have not specifically included the effect of weld-metal undermatching or overmatching. The assumption has been that the weld overmatches, and this has been implicitly true for the lower grades of pipeline steels that have been welded with conventional pipeline welding consumables.

Strength properties of the pipe and the weld, however, are variable and some level of mismatch will occur in nearly all welds. Undermatching is likely when pipe from the high end of the strength distribution is joined with welds from the low end of the distribution.

With the now-common use of pipe materials of Grade 483 and 550, it is not easy to ensure overmatching. Before specifying overmatching, the designer should carefully consider such detrimental effects as the following:

• Weldability decreases as strength increases, which will result in the potential for more defects.
• These defects could be of the more injurious planar type, including hydrogen cracks.
• Toughness usually decreases with increasing strength.

It is important to note that for most pipeline designs, slight undermatching in girth welds is of no concern because defects are small, the applied stress is elastic, the welds have good toughness, and there is added benefit from weld reinforcement.

The stress field surrounding a defect governs fracture and collapse behavior. Because mismatch can significantly affect the development of plastic strains around defects, it must be taken into account when determining tolerable defect sizes.

Because plasticity develops easier in weaker materials, conservative results will be obtained from most fracture and collapse analyses if the yield strength of the weaker material is used in the calculations. The fracture-toughness properties which should be used are those of the material in which the defect is located.

These slight modifications, when applied to existing assessment methods, should restore their original intended levels of conservatism while accounting for the effects of mismatch. A generalized methodology is shown in Fig. 4 [123,008 bytes].

The key objective of the acceptance criteria should be maintaining cost-effective designs while achieving fitness-for service.

Fracture control

As a result, the choice of appropriate measures for controlling fracture initiation and fracture propagation for high-strength pipelines assumes an important role.

Equations developed by Battelle for the American Gas Association⁸ are widely used to evaluate the resistance of line pipe to initiation of fracture from axial flaws. Application of these equations shows that the critical flaw size decreases with increasing yield strength; that is, the toughness required to prevent crack initiation from a given flaw size increases with increasing strength.

The Battelle equations are purely deterministic. Work is being carried out, however, to develop a probabilistic approach, based on a failure-assessment diagram, thus incorporating both fracture and collapse. Either approach will show that, for modern high-strength steels, it is possible to obtain adequate toughness to prevent fracture initiation.

The current approach to determine the required fracture-arrest toughness relies on the correlation between Charpy V-notch energy and full-scale test behavior. Several semiempirical relationships exist to predict the required toughness for the prevention of fracture propagation.

While these different approaches are not necessarily in good agreement, the trend is clear that a higher level of toughness is required with increasing yield strength for equivalent design conditions. Recent work on Grade 550 material has shown that some fractures propagate at Charpy levels higher than that of the Battelle equation.

On this basis, it was recommended that the energy predicted by the Battelle equation for fracture arrest should be increased by 30% for higher-strength steel. Recently, a new approach utilizing the crack-tip-opening angle (CTOA) as a characterizing parameter has been proposed for high-strength steels.⁹

With this approach, the fracture resistance of the material (CTOAc) is compared to the driving force of the pressurized gas for a given pipeline design. This small-scale test would appear to give improved correlations for steels up to Grade 550.

Other work is also studying the use of drop-weight tear tests to develop ductile fracture-arrest methods for modern pipe steels and design.¹⁰

In either case, as the gas properties change and the design pressure increases, the aspect of fracture arrest becomes more important, and the understanding of the characterizing parameters become critical.

Eastern Alberta system

The mechanized gas-metal arc welding procedures employed for main line welding were identical to those used for Grade 483; that no welding problems were encountered could be attributed to the process, the procedure, or the higher strength material being welded.

The mechanized GMAW system used consisted of an internal welding machine, one unit of two tractors (one welding "shack") for the hot pass, four shacks for fill passes, and four shacks for the cap pass. Welding productivity was 110 joints/day with a repair rate of 6%, which is consistent with experience of welding Grade 483 in this diameter.

Tie-ins were completed with a combination of cellulosic SMAW (E55010G) for the root and hot pass, with 100° C. preheat, followed by self-shielded FCAW for all remaining passes.

The particular self-shielded consumable selected was optimized in terms of deposit strength and toughness by the manufacturer for application to Grade 550 pipe, and the welds produced consistently met yield-strength requirements and exceeded the toughness requirements at the -5° C. design temperature.

In addition, experience during construction demonstrated that the tie-in welds were completed some 40% faster with self-shielded FCAW welding than equivalent welds made with conventional cellulosic electrodes throughout. Low-hydrogen, vertical-down SMAW made with E69018G electrodes was used for any repairs carried out to the mechanized gas-metal arc welds.

Central main line loop

The welding procedure was identical to that used on the 1994 project with additional fill passes for welds in the 16-mm W.T. pipe. The spread of mechanized welding equipment involved an additional fill-pass shack; 130 joints/day were achieved at a repair rate of 7%.

Tie-in and repair welding was as stated previously.

Eastern main line loop

One welding shack would complete the root pass, and three additional shacks would each complete the remaining hot, fill, and cap passes of a weld. Production rates of some 70 welds/day with repair rates of around 5% were achieved.

This configuration was the most economical for the length and location of this particular construction project. Low-hydrogen, vertical-down SMAW with cellulosic root and hot passes were used for tie-ins and repairs.

The future

Manual welding processes and consumables are being investigated to give the contractor a wider choice of welding methods.

TCPL will look at higher-strength steels for specific applications. Japanese and European manufacturers have produced trial quantities of Grade 690 pipe material and are actively researching higher grades.

Even though higher-strength pipe may be available, however, the other elements of a piping system, fittings, are not, as yet, easy to obtain at the Grade 550 strength level.

References

1. Somerville, F.S., Slimmon, T.C., "Property Requirements for Pipelines," Materials Engineering in the Arctic, ASM, 1977.
2. Shelton, E., Rothwell, A.B., and Coote, R.I., "Steel Requirements for Current and Future Canadian Gas Pipeline Systems," Metals Technology, 1983.
3. Dorling, D.V., and Rothwell, A.B., "Field Welding Processes for Pipeline Construction," Pipeline Technology Conference, Ostende, October 1990.
4. Coote, R.I., Glover, A.G., Pick, R.J., and Burns, D.J., "Alternative Girth Weld Acceptance Standards in the Canadian Pipeline Code," 3rd International Conference Welding and Performance of Pipelines, TWI, London, November 1986.
5. Glover, A.G., Hodgkinson, D., and Dorling, D.V., "The Application of Mechanized Ultrasonic Inspection and Alternative Acceptance Criteria to Pipeline Girth Welds," Pipeline Technology Conference, Ostende, October 1990.
6. Kostic, M.M., Gedeon, S.A., Bowker, J.T., and Dorling, D.V., "Development and Production of X80 (550 MPa) Gas Transmission Linepipe," proceedings of the 2nd International Pipeline Technology Conference, Ostende, 1995.
7. Horsley, D.J., and Glover, A.G., "Girth Weld Strength Under-Matching in High Pressure Natural Gas Pipelines," Mis-Matching of Interfaces and Welds, eds. K.-H. Schwalbe and M. Kocak, 1997, GKSS Research Center Publications, Geesthacht, FRG.
8. Maxey, W.A., "Fracture Initiation, Propagation and Arrest," 5th Symposium on Line Pipe Research, AGA, Catalog No. L30174, 1974.
9. Popelar, C.H., "Ductile Fracture Propagation Model Development," Private communication, on-going PRCI Research, 1997.
10. Wilkowski, G.M., and Mihell, J., "Ductile Fracture Arrest Methodology for Current and Future Grades of Line Pipe Steels," CIM Conference, August 1998, Calgary.

The Authors

Alan Glover is an engineering specialist with Trans Canada PipeLines' quality, standards, and technology group. He holds a BS (honors) in metallurgy from the University of Leeds, U.K., and a PhD in fracture and fracture mechanics. He is a registered professional engineer in Alberta.

David Horsley is a senior engineer in TransCanada PipeLines' quality, standards, and technology group. He holds BS and MEngr degrees in mechanical engineering from the University of Calgary and is a registered Professional Engineer in Alberta.

David Dorling is an engineering specialist with TransCanada PipeLines' quality, standards, and technology group. He holds a BS (honors) in metallurgy from the University of Manchester and a PhD in welding engineering from the University of Cranfield. He is an adjunct professor in the department of chemical and materials engineering at the University of Alberta and a registered professional engineer in Alberta.

Copyright 1998 Oil & Gas Journal. All Rights Reserved.